Header self-links gone?

Header self-links gone?

[Nick Moffitt]

I resurrected a rather hairy old project from a year ago, and rebuilt it
using pandoc 1.12.4.2~dfsg-1build4 and well, here's my questions about
what's changed in the meantime:

    1: It used to be that with section-divs and toc, I'd still get
       self-links on section headers.  So like:

        ## Foo
       would make
        <h2><a href="#foo">Foo</a></h2>

       Is that just gone, or is there some extension I can turn on to
       get that back?  I was using those links for javascript and css
       behaviours.

    2: I'm building with `--standalone --self-contained` and that used
       to make my `-c local.css` inline into the single file, but if I
       explicitly switch on any options at the end of my html5 output
       type, it seems to switch to external linkage.  I may be missing
       some actual behaviour here, as I've not tested this to
       destruction.

Re: Header self-links gone?

[John MacFarlane]

+++ Nick Moffitt [Feb 27 15 14:10 ]:
>I resurrected a rather hairy old project from a year ago, and rebuilt it
>using pandoc 1.12.4.2~dfsg-1build4 and well, here's my questions about
>what's changed in the meantime:
>
>    1: It used to be that with section-divs and toc, I'd still get
>       self-links on section headers.  So like:
>
>        ## Foo
>       would make
>        <h2><a href="#foo">Foo</a></h2>
>
>       Is that just gone, or is there some extension I can turn on to
>       get that back?  I was using those links for javascript and css
>       behaviours.

The automatic IDs used to be added by the HTML writer.  (And certain
other writers.)  Now they are added by the readers (if
`auto_identifiers` is set.

So if you're generating a Pandoc programatically, you'll need to
add identifiers to the header elements yourself.  You can use
uniqueIdent from Text.Pandoc.Shared to do this.

>    2: I'm building with `--standalone --self-contained` and that used
>       to make my `-c local.css` inline into the single file, but if I
>       explicitly switch on any options at the end of my html5 output
>       type, it seems to switch to external linkage.  I may be missing
>       some actual behaviour here, as I've not tested this to
>       destruction.

Not sure here.  Can you give a concrete example that I can use to
reproduce?

Re: Header self-links gone?

[Nick Moffitt]

John MacFarlane:
> The automatic IDs used to be added by the HTML writer.  (And certain
> other writers.)  Now they are added by the readers (if
> `auto_identifiers` is set.
> 
> So if you're generating a Pandoc programatically, you'll need to add
> identifiers to the header elements yourself.  You can use uniqueIdent
> from Text.Pandoc.Shared to do this.

So if I'm using the pandoc command-line program, can I just `-f
markdown+auto_identifiers`?  That didn't seem to work for me.

> >   2: I'm building with `--standalone --self-contained` and that used
> >      to make my `-c local.css` inline into the single file, but if I
> >      explicitly switch on any options at the end of my html5 output
> >      type, it seems to switch to external linkage.  I may be missing
> >      some actual behaviour here, as I've not tested this to
> >      destruction.
> 
> Not sure here.  Can you give a concrete example that I can use to
> reproduce?

My Makefile does this to get single-file self-contained documents:

```
%.html: %.md template.html5 footer.html5 mystyle.css
	pandoc -t html5 --ascii --toc --toc-depth=2 -c mystyle.css --template=template.html5 --include-after=footer.html5 --section-divs --standalone --self-contained --smart -o $@ $<
```

If I change `-t html5` to `-t html5+foo`, then `mystyle.css` will become
a link instead of a blob of base64 goo.  I think I'm misunderstanding
writer extensions, though, and just confusing it entirely.  Perhaps it's
because of my custom template?  It was a surprising symptom to me, in
any case.

Re: Header self-links gone?

[dev.crst-Re5JQEeQqe8AvxtiuMwx3w]

[-- Attachment #1.1: Type: text/plain, Size: 1967 bytes --]

Nick Moffitt:
>
>     1: It used to be that with section-divs and toc, I'd still get 
>        self-links on section headers.  So like: 
>
>         ## Foo 
>        would make 
>         <h2><a href="#foo">Foo</a></h2> 
>

 A few days ago I upgraded Linux Mint and was also wondering why this 
"Self-Link" feature wasn't working anymore. After some naive investigation 
I found the answer in the changelog <http://pandoc.org/changelog.txt> down 
at "pandoc (1.12.3)":

  * HTML writer:

    + With `--toc`, headers no longer link to themselves (#1081).

At #1081 <https://github.com/jgm/pandoc/issues/1081> there is the reason 
for this.

I also missed this feature and my workaround I can live with is a little 
javascript in the html-template's header:

<script>
    function headerLink() {
        for (var level = 1; level <= 6; level++) {
            var header = document.getElementsByTagName("h" + level);
            var i;
            for (i = 0; i < header.length; i++) {
                header[i].innerHTML =
                    '<a href="#'
                    + header[i].id
                    + '">'
                    + header[i].innerHTML + '</a>';
            }
        }
    }
</script>

Now just change
<body>
to
<body onload="headerLink()">

is there some extension I can turn on to get that back?
>

Yeah that'd be *great* <https://memegenerator.net/Yeah-ThatD-Be-Great>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/ac25089d-f112-4795-8524-ac705b521fcc%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 11733 bytes --]

Re: Header self-links gone?

[Nick Moffitt]

dev.crst-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org:
> At #1081 <https://github.com/jgm/pandoc/issues/1081> there is the reason 
> for this.
> 
> I also missed this feature and my workaround I can live with is a little 
> javascript in the html-template's header:

Oh dear.  Part of why I use pandoc is to get AWAY from JavaScript!

This is extremely disheartening.

Re: Header self-links gone?

[John MacFarlane]

We could consider adding an option to add in the selflinks
(for HTML only, presumably).

Re: Header self-links gone?

[Christoph Strasser]

[-- Attachment #1.1: Type: text/plain, Size: 910 bytes --]

Made my day!

For my purpose I currently use pandoc to just convert between markdown and 
html. Therefore, this option would solve my only (JavaScript-)workaround so 
far. 

After upgrading Linux Mint I took the chance to install the latest version 
(1.18) manually and it is really a great pleasure to see how the 
development advances!

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/6b83c3f3-9e7d-4d22-9aa2-0ad6fafdc0f5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 1338 bytes --]

Re: Header self-links gone?

[BP Jonsson]

Den 2016-11-23 kl. 12:22, skrev Nick Moffitt:
> dev.crst-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org:
>> At #1081 <https://github.com/jgm/pandoc/issues/1081> there is the reason 
>> for this.
>>
>> I also missed this feature and my workaround I can live with is a little 
>> javascript in the html-template's header:
> 
> Oh dear.  Part of why I use pandoc is to get AWAY from JavaScript!
> 
> This is extremely disheartening.
> 

If more JavaScript is odious to you a filter which modifies the static content may seem a little better. Here are both a HTML (output) filter

    pandoc -t html5 input.md | perl html-self-links.pl >output.html

and a pandoc filter 

    pandoc -F pandoc-self-links.pl -t html5 input.md -o output.html
 
doing the same thing.

<https://gist.github.com/bpj/963fddaea671c6fd4d2ebe66b27e7960>

I don't know which is faster so you may want to time them if it is important. The html filter has the advantage that you can modify existing HTML files with it.

/bpj

Re: Header self-links gone?

[Christoph Strasser]

[-- Attachment #1.1: Type: text/plain, Size: 1927 bytes --]

On Thursday, November 24, 2016 at 12:24:45 AM UTC+1, BP Jonsson wrote:

> If more JavaScript is odious to you a filter which modifies the static 
> content may seem a little better. Here are both a HTML (output) filter 
>
>     pandoc -t html5 input.md | perl html-self-links.pl >output.html 
>
> and a pandoc filter 
>
>     pandoc -F pandoc-self-links.pl -t html5 input.md -o output.html 
>   
> doing the same thing. 
>
> <https://gist.github.com/bpj/963fddaea671c6fd4d2ebe66b27e7960> 
>
> I don't know which is faster so you may want to time them if it is 
> important. The html filter has the advantage that you can modify existing 
> HTML files with it. 
>
> /bpj 
>
>
*Thank you!*

The big advantages of using a filter is that also textbased browsers enjoy 
selflinks plus computing has just to be done once. Here are the test 
results with MANUAL.txt <http://pandoc.org/demo/MANUAL.txt> as input file:

*Without filter*

% time pandoc -s -t html5 MANUAL.txt -o output.html

real    0m0.366s
user    0m0.368s
sys     0m0.000s


*html-self-links.pl*

% time { pandoc -s -t html5 MANUAL.txt | perl html-self-links.pl > output.html 
; }

real    0m0.516s
user    0m0.556s
sys     0m0.012s


*pandoc-self-links.pl*

% time { pandoc -s -F pandoc-self-links.pl -t html5 MANUAL.txt -o output.html 
; }

real    0m2.362s
user    0m2.360s
sys     0m0.020s


-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/bb4ed5a7-c81e-420b-bdc6-74a3b675a653%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #1.2: Type: text/html, Size: 10478 bytes --]

Re: Header self-links gone?

[Christoph Strasser]

[-- Attachment #1: Type: text/plain, Size: 1896 bytes --]

Well, I was just curious and tried as bedtime candy to write the 
html-self-links.pl 
<https://gist.github.com/bpj/963fddaea671c6fd4d2ebe66b27e7960#file-html-self-links-pl> 
in |bash|. While doing so I noticed that the new |--section-divs| option 
is not covered as with the javascript above. The pandoc-self-links.pl 
<https://gist.github.com/bpj/963fddaea671c6fd4d2ebe66b27e7960#file-pandoc-self-links-pl> 
does it, but needs it's time. Since I want to handle pandoc via a 
bash-script with no other dependencies, I think when already "taping", 
in my case piping to a function could be the best choice.

You have to know that I am a |perl| virgin which is fleeing from it like 
Syrinx from Pan, although every time my mojo rises when it is near... I 
guess one day it will get me anyways.

html-selflink 
<https://github.com/devcrst/bash-chunk/blob/master/pandoc/html-selflink>

|% time { pandoc -s -N -t html5 --section-divs --toc MANUAL.txt | bash 
html-selflink > output.html ; } real 0m0.555s user 0m0.512s sys 0m0.048s|

The best solution undoubtedly would be an option, maybe like:

|--header-links=none|self|toc|

where |toc| links back to the table of content entry? But for me it is 
obvious that this has to come if at all when time allows. It seems that 
not many are missing this out there.

BTW, You are my heroes!

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/6049b086-bbeb-642c-2b42-55b38e8d107e%40gmail.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #2: Type: text/html, Size: 2728 bytes --]

Re: Header self-links gone?

[BP Jonsson]

[-- Attachment #1: Type: text/plain, Size: 4061 bytes --]

I don't know how your bash script works but I guess it uses a sed regex to
find and modify headings, which is rather fragile. Both my Perl scripts are
actually wrappers around more robust parsers: the one uses Mojo::HTML,
which uses a regex to parse HTML but a monster regex coupled with quite
some error checking. The other actually uses pandoc as its parser, and they
use somewhat sophisticated libraries (modules in Perl parlance) to traverse
the parse tree, Mojo::DOM/Mojo::CSS in the one case and Pandoc::Elements in
the other, which makes things somewhat more robust. There is simply less
that can go wrong that way (read: runaway matches! :-) I did do quite a lot
of HTML editing with simplistic regexes with frequently disastrous results
in the old days[^1] You can easily make html-self-links.pl executable,
place it in your PATH and then invoke it from a bash script or Makefile
just as you invoke pandoc. That way you don't need to proofread your HTML
for runaway substitutions.

[^1]: Back in the early to mid nineties here in Sweden web development was
an easy way for a student with some computing skills to make some extra
money. You could pull off quite legitimate web sites entirely without
JavaScript, which still was in its infancy in those days.

Den 25 nov 2016 22:10 skrev "Christoph Strasser" <dev.crst-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>:

> Well, I was just curious and tried as bedtime candy to write the
> html-self-links.pl
> <https://gist.github.com/bpj/963fddaea671c6fd4d2ebe66b27e7960#file-html-self-links-pl>
> in bash. While doing so I noticed that the new --section-divs option is
> not covered as with the javascript above. The pandoc-self-links.pl
> <https://gist.github.com/bpj/963fddaea671c6fd4d2ebe66b27e7960#file-pandoc-self-links-pl>
> does it, but needs it's time. Since I want to handle pandoc via a
> bash-script with no other dependencies, I think when already "taping", in
> my case piping to a function could be the best choice.
>
> You have to know that I am a perl virgin which is fleeing from it like
> Syrinx from Pan, although every time my mojo rises when it is near... I
> guess one day it will get me anyways.
>
> html-selflink
> <https://github.com/devcrst/bash-chunk/blob/master/pandoc/html-selflink>
>
> % time { pandoc -s -N -t html5 --section-divs --toc  MANUAL.txt | bash html-selflink > output.html ; }
>
> real    0m0.555s
> user    0m0.512s
> sys     0m0.048s
>
> The best solution undoubtedly would be an option, maybe like:
>
> --header-links=none|self|toc
>
> where toc links back to the table of content entry? But for me it is
> obvious that this has to come if at all when time allows. It seems that not
> many are missing this out there.
>
> BTW, You are my heroes!
>
> --
> You received this message because you are subscribed to the Google Groups
> "pandoc-discuss" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To view this discussion on the web visit https://groups.google.com/d/
> msgid/pandoc-discuss/6049b086-bbeb-642c-2b42-55b38e8d107e%40gmail.com
> <https://groups.google.com/d/msgid/pandoc-discuss/6049b086-bbeb-642c-2b42-55b38e8d107e%40gmail.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups "pandoc-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pandoc-discuss+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To post to this group, send email to pandoc-discuss-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To view this discussion on the web visit https://groups.google.com/d/msgid/pandoc-discuss/CAFC_yuRrspsV2WFLa%3Dj%2BjYM9sVJnh4hGZY%2Bm%2BmRQsCfdLST34w%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[-- Attachment #2: Type: text/html, Size: 5447 bytes --]